Multi-Level Data Distribution with Raft Replication

When you select Multi-Level data distribution with Raft replication, continue the configuration of the following options:

Number of data spaces: Select the number of data spaces you need for your Multi-Level data distribution topology.

Configure deployment options

In the table, configure the deployment options for each data space.

Data space: Enter names for the data spaces.
Deployment options: Select a Single-Raft Cluster or Multi-Raft Cluster topology.
- Single-Raft Cluster: You will deploy a single read-write Raft cluster.
- Multi-Raft Cluster: You will deploy a read-write Raft cluster and one or two read-only Raft clusters that are replicas of the read-write cluster.
Number of clusters: With Single-Raft Cluster option selected, you will have 1 number of clusters. With Multi-Cluster Raft option selected you can select 2 or 3 clusters.
Replication units: Select the number of replication units. A Globally Distributed Autonomous AI Database contains multiple replication units. A replication unit is a set of chunks that have the same replication topology.
Chunks: you can set the number of chunks per data space.

Select Edit in the Actions menu (three dots) to edit Raft cluster details in each data space.

Raft cluster name: You can rename the Raft cluster here.
Number of nodes: Indicate the number of nodes (databases) to deploy in the Raft cluster.
Mode: Applies to Multi-Raft Cluster deployment option. Select whether the Raft cluster will be read-write or read-only.
Witness: (Optional) Enable this option to designate one follower of each replication unit as a witness. A witness retains the Raft consensus module but only stores a portion of the Raft log and does not contain a table partition. The witness replica actively participates in nearly every aspect of the consensus protocol.

Use the Show map toggle to select between the map and list view to configure the nodes.

Map view: The map view filters and shows you the available Exadata clusters where nodes can be deployed. To configure nodes using the map, select an available region, select Configure nodes.
Table view: With Show map switched off you can configure the nodes in a tabular format. To configure each node, select Edit in the Actions menu (three dots).

Whether in Map or Table view, you configure the following options for the nodes.

Region: Select the OCI region where you would like to host your node.
Data space: Select the data space the node should belong to.
VM cluster: Select a cluster available in the selected region.

Note: It is recommended that you use one VM cluster per node.
ECPU count: Enter the number of ECPU cores to enable for each node. Specify the number of ECPUs as an integer. Available cores are subject to your tenancy’s service limits.

You must enter a minimum of 16 ECPUs per node.

ECPUs are based on the number of cores, elastically allocated, from the shared pool of Exadata database servers and storage servers. Aggregated ECPU consumption on a given cluster is 1.5 times the ECPU count.

Note that a number of ECPUs are consumed in overhead and are not available to the nodes.

See Oracle Cloud Infrastructure Documentation for more information.
Auto scaling: Enable automatic scaling based on workload per node. This value is passed on to the Autonomous AI Database so that it can manage ECPU auto scaling.

See Enable or Disable Auto Scaling of an Autonomous AI Database on Dedicated Exadata Infrastructure for details.
Storage: GB of storage to allocate to the node (database)

With nodes configured you can return to Catalog configuration.